That’s Trend Micro, a leading antivirus company. Trend Micro is quite instrumental also in counter-disinformation, because they’re of the first cybersecurity companies that published a disinformation analysis report a few years ago.
Then they also developed a chat bot, a Line bot, that scans each incoming messages. Like if you have a WhatsApp group, actually a Line group in Taiwan, because people in Taiwan use Line more. You invite that bot into your chatroom.
That bot scans each incoming messages as an antivirus program would do, and compare it with a crowdsourced, what we call Cofacts, collaborative fact-checking, ecosystem. So that if the message being shared in the group is a disinformation package, that bot, within split second, responds saying, “This has been clarified, and this is not the case.”
If people receive a disinformation, go to sleep, and wake up to see the clarification, that’s of little use, because the disinformation already framed their understanding. It’s already written in their long-term memory.
It already maybe reinforce their thought patterns. Within split second, then it is an idea in, very old idea in journalism, called a balance in perspective and report. If you have one source, you need to check the other source.
The Trend Micro bot built this kind of balance, again, in a very unlikely place, which is group chat. That is an important contribution.
Sure, and my colleague, Joel, can send you what we call a norm package, which is a counter-disinformation, self-regulating agreement, a kind of a pact between the likes of Google, Facebook, Yahoo, Line, PTT, and so on.
What they have broadly agreed to the norm is that one of not only transparency and accountability, but also life-long education and empowerment. FB, for example, partnered with the Hondao Elderly Care Foundation to make sure that the digital competency information packages is not just targeting the young people, but actually can be localized for very old people.
Like people in their 80s and so on, who perhaps don’t like being corrected by the Trend Micro bot in their family chatrooms all the time. They also want to be a contributor to correct their grandchildren’s messages. [laughs]
They also have incentive to learn this kind of media literacy. Our Control Yuan, which is a separate branch in the government, establish a norm where all the campaign donation and expense is transparent, down to the raw data level.
Not just the statistics, but available for independent investigative journalists, as well as data scientists, to draw conclusions based on political contribution. This is similar to the US of honest advertisement when it comes to campaign financing, but in real time, and in raw data form.
We just told the major social media companies, saying that, “This is the norm in Taiwan, that we expect democracy to work like this.” We see clearly from the Control Yuan data that there are certain expenses that are for precision targeting in your social media platform that are neither filed as campaign donation, nor filed as campaign expense.
These are filed as people friendly to the candidate, voluntarily buying, huge number of money, into precision targeting. We think that’s violating our norm. You have two choices. First, you can agree to open up again in real time whatever precision targeting terms and criteria is being put forward in your platform.
That’s what Facebook have done in the ads library, so that if there is a candidate that use this hyper-precision targeting, spreading disinformation to discourage people to vote, they can be called out within one hour and face social sanction.
That’s FB, what they have done. They set up a war room to ensure a rapid response to such divisive or counterfactual advertisements. That’s what FB did. Google, for example, along with Twitter, simply said, “OK, maybe we don’t run political and social advertisements during your election.” That’s fair, too.
Maybe one more question.
Sure, of course.
That’s a great question. I think a clear legal definition of disinformation is the first step, because without such a disinformation definition, it’s very easy to be politicized as a term. We mean disinformation as intentional untruth, intended to harm the public, not just an image of a government, which may be just good journalism.
This intentional public harm, all built upon existing legal concepts of the pre-digital media legal system, so the court can very predictably make judgments based on these criteria. We’re not inventing novel legal concepts for disinformation.
We’re just saying whatever previous acts that discourage people or penalize people for spreading misinformation when – SARS, for example – an outbreak occurs, which it actually is doing in our nearby jurisdiction.
In any case, whenever this kind of issues happens, we understand that democracy builds upon the health. Literally, the health of people. Democracy should prevent this kind of disinformation from spreading.
This is not about free speech if somebody, intentionally through sharing of data and information that are simply they know it’s not true, that endanger other people’s lives. That, of course, is clearly outside freedom of expression. That’s the legal concept.
The other thing is that we say to the journalistic community that we are partners, that we’re doing this to say that we’re encouraging your fact-checking efforts by first, never call ourself, the administration, the fact-checkers.
We’re just people who provide a real time clarification, sometime in a very funny way, for the journalists to work with. The legitimacy is in the social sector and in the journalism business. That is the first thing.
The minister’s words never sits above a journalist’s word, and we never take away journalistic freedom by, for example, issuing takedowns for journalistic output. According to the human right society, Civicus Monitor, we’re the only jurisdiction in Asia that completely implements this stance.
Many other jurisdictions in Asia implemented a certain amount of administrative override to journalism and free expression because of disinformation crisis. Our way of innovating, I think, commits to the free and open value by issuing, instead of takedown, real time clarification, like the Trend Micro bot.
Everybody in the chat group, instead of thinking that, “Oh, there is messaged disappeared by censors,” they would instead learn about something, about the journalistic value of fact-checking and so on.
The other thing that you might or might not want to include is that we refrain from using the word “fake news,” because news and journalism in Mandarin in the same word. There’s no way to say the F-word without offending journalists. Both of my parents are journalists, so out of filial piety, I just cannot say the F-word. [laughs]
Thank you.
Welcome.
The digital dialogues.
That’s right.
The other way around. How Taiwan can help.
As a matter of fact, when you look at the Taiwan FactCheck Center, tfc.taiwan.org.tw which is an independent nonpartisan, nongovernment funded, social sector organization that takes a look at trending disinformation and do attribution work, investigative work on it, it’s very clear that there were several messages that originated from the PRC.
For example, during the anti-ELAB protests, there was a trending message in the Taiwan social media landscape that tried to portray the Hong Kong protesting people as so-called rioters, that pays alleged $20 million to murder police.
That is of course a gross misrepresentation of their mission, but what this is trying to do is obviously try to lessen the effect that the anti-ELAB movement has on the Taiwan presidential election.
You see it in actually more trending in the Taiwanese social media than in Hong Kong, because they would easily see through this disinformation package. It’s certainly not targeted at Hong Kong audience.
Now the TFCC actually attributed original wording of this already heavily remixed, and illustrated version of this disinformation back to the Weibo account of 中央政法委长安剑, the PRC central political and law unit’s Weibo account, and they provided the original fabrication that misuses Reuters photo, and try to paint it as a kind of teenage protestor to engage in violent activity just to buy some iPhones and things like that.
The second level remixers then create supporting narratives on that. It is clearly attributed as a PRC action.
The one that I just cited, which is number 204, was published by the TFCC in November 15, so already quite close to the election. I would say that it was part of the pre-election disinformation landscape. After the election as you pointed out, there’s a flurry of disinformation packages that tries to invalidate the election result, and meanwhile, try to sow discord between Taiwanese people and the US.
For example, there was a disinformation package that says when Tsai Ing-wen was printing the voting ballot, she instructed the CUC to use a special invisible ink developed by the CIA so that no matter who you vote for, you will end up voting to Tsai. This is of course against the physical law [laughs] the law of physics.
The clarification was mostly interviewing chemistry teachers, but the more importantly, we can see the intention of this disinformation. It’s not only to sow discord in the democratic process, but try to point a finger at the US.
What really this is trying to do, is try to make people feel less secure about the democratic process, and thereby engage in a more authoritarian way, and make them more susceptible of future disinformation packages. That’s my take, what is doing postelection.
A lot of disinformation don’t quite spread by itself. It spreads only through people who want to remix it, to mobilize people in outrage. That is main emotion behind most disinformation packages.
If you hear about alleged young people being paid to murder police, or if you hear about CIA supplying invisible ink, people will feel angry and helpless, a little bit, which is a kind of paralyzing emotion. They get out of this emotion by clicking share, and write something motivating for their social groups. Then it carries their credentials for their friend circles.
That’s how disinformation spread by provoking outrage. It’s kind of like memetic virus outbreak. One of the way that work with the fact checkers, for example TFCC, is not only to provide them with real-time clarification messages. For example, on exactly how the ballot’s printed. It’s printed by each municipality and cities, and the CUC doesn’t actually print anything.
That’s the factual response. It’s also equally important that we frame our clarification messages in a way that is humorous. This idea of a timely response from rumors and clarification, which require the kind of deadline is two hour after each trending disinformation is detected. Many ministries can now do so in 60 minutes.
What they do, is basically provide a clarification to a trending rumor by using humor, especially humor that corresponds to existing memetic, what we call “gung” here, which is memes, Internet memes, to ensure that people find it genuinely funny, hilarious even. When people laugh about the humor, the anger is vented into humor.
They would also organically share humorous messages, so the clarification become viral more so than the disinformation, and because anyone who view this humorous repackage of clarification and laughed about it, becomes immune to outrage, because humor and outrage are two outlets of anger and they are mutually exclusive. Then people become inoculated against this kind of outrage.
